backend-ai-tools
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXFILTRATION] (SAFE): The provided code examples demonstrate standard database CRUD operations (e.g., Resource.save, Project.find). These operations are consistent with the skill's stated purpose of creating backend tools and do not involve unauthorized data movement or hardcoded credentials.
- [REMOTE_CODE_EXECUTION] (SAFE): No patterns for remote code execution or piped shell commands (e.g., curl|bash) were detected.
- [PROMPT_INJECTION] (SAFE): The skill does not contain instructions designed to bypass AI safety guardrails or override system instructions.
- [Indirect Prompt Injection] (LOW): The skill creates surfaces for ingesting untrusted data through tool inputs (e.g., in
apps/backend/src/ai/tools/project.ts). However, it mitigates risk by documenting and enforcing strict Zod schema validation for all inputs. Capability inventory includes database read/write access. No unsafe interpolation into larger LLM prompts is demonstrated in the skill logic itself.
Audit Metadata